A research on capability analysis methods used in cases of non normally distributed process data

Process capability analysis is used to measure the ability of a production process to meet the tolerances specified for the quality characteristics of the manufactured products. Process capability indices are the most important tools used to measure the ability of the processes.Early proposed process capability indices in the literature work under the assumptions that the process data is normally distributed, the tolerances of quality characteristics are symmetric and the process is under control. In further studies it's seen that various capability indices are proposed in cases where (i) the process data is distributed normally and the tolerances are asymmetric, (ii) the process data is asymetrically distributed and the tolerances are symmetric.Apart from other studies in the literature, in this thesis, a new process capability index is proposed for the cases where the tolerances are asymmetric and the distribution of the process data is not normal. Pearson distribution family is used in order to find out what an asymmetric distribution desired to beproduced by the production process should be and to determine how close are the actual data produced by production process to this desired distribution. The performance of the proposed index in various cases is examined with examples. In addition, an application of the proposed process capability index in a valve producing factory has been made and the results have been evaluatedKeywords: Asymmetric Tolerance, Asymmetric Distribution, Process Capability Indices, Pearson Distibution Family.